4.1. Answer to Research Question 1
Table 2 presents evidence to determine whether there are signs of gender bias in the perception of the impacts arising from the event. The analysis compares responses separately by gender rather than drawing conclusions from the overall sample means; therefore, the unequal proportion of women (57.84%) and men (42.16%) in the sample does not account for the observed direction of the gender differences. As shown in
Table 2, descriptive statistics and between-group comparisons are reported separately for women and men. Specifically, women tend to give higher scores than men to items relating to the perception of inclusive impact, sporting impact and the promotion of the city’s image. In this regard, only IDT8 (promotes sport amongst people with disabilities) shows no statistically significant differences. Furthermore, virtually all the medians for women are at the maximum value on the scale and 1 point above the corresponding value for men. This is not the case, however, for IDT4 (visitors to Córdoba will find this event appealing). When analysing the standard deviations, it can be seen that the greatest consensus in responses is found in the items relating to the perceived impact on inclusion, sport and the promotion of the city’s image, and this applies only to women, whilst for men there is greater disparity of opinion.
Perceptions of economic impact show statistically significant differences across all items (ECO1 to ECO4). The medians for women (6 points) are also 1 point higher than those for men (5 points). Furthermore, the items relating to the perception of social impact (SOC1 to SOC6) also show a gender-based bias in perception, with the exception of SOC6 (‘This event positions Córdoba as a safe city’), although women’s mean score is also higher than men’s. In the social sphere, however, there is not such a significant difference between women’s and men’s mean scores.
Finally, perceptions of environmental impacts or urban disturbances caused during the event do not show any statistically significant differences by gender across any of the items. Despite this, the standard deviations are the highest in the questionnaire for both men and women. This suggests the existence of a perceptual bias stemming from other factors and calls for further exploration of the analysis proposed by this research. Therefore, in response to RQ1, the results provide evidence of gender differences in residents’ perceptions of event impacts, with women tending to give higher ratings than men regarding the promotion of an inclusive and sporting image of Córdoba, as well as perceived economic and social impact. However, no statistically significant gender differences were found in perceptions of environmental impact or urban disruption.
Despite the statistical significance observed for several items, all significant differences showed small effect sizes, ranging from r = 0.11 to r = 0.24, with the largest effect observed for IDT7 (r = 0.24). Conversely, non-significant comparisons showed trivial effect sizes, ranging from r = 0.01 to r = 0.08. Thus, although the results indicate a consistent pattern of gender differences in several perceived impacts, their magnitude should be interpreted as small.
4.2. Answer to Research Question 2
The results of the EFA (
Table 3) identified a structure comprising four distinct dimensions of perceived impact: inclusive-sporting impact (IDT), environmental impact (ENV), social impact (SOC) and economic impact (ECO). Each of the identified dimensions comprised at least four items, all of which had factor loadings greater than 0.50 and were clearly associated with a single factor, with no issues of cross-saturation. The eigenvalues of the four extracted factors were greater than 1, meeting the standard criteria for factor retention. Overall, the factor solution explained 72.72% of the total variance, indicating the model’s high explanatory power. In particular, the inclusive-sporting impact (IDT) factor had the greatest explanatory weight, followed by the environmental, social and economic dimensions. The internal reliability of the dimensions was adequate, with high values for both Cronbach’s alpha (α between 0.90 and 0.92) and McDonald’s omega (ω between 0.90 and 0.93), which supports the internal consistency of the identified scales. These results highlight the emergence of a specific dimension linked to the perception of the promotion of the city’s sporting and inclusive image, distinct from the economic, social and environmental dimensions traditionally associated with the Triple Bottom Line approach, thereby providing an affirmative answer to RQ2.
These results demonstrate the existence of a framework for perceiving impact based on the traditional TBL approach (perception of economic, social and environmental impact), alongside a dimension relating to the promotion of the city’s sporting and tourism image, thereby providing an affirmative answer to RQ2.
4.3. Answer to Research Question 3
The CFA sought to identify a structure of perceived impact and support consistent with the SET (
Table 4). These latent constructs were: inclusive-sporting impact (IDT), environmental impact (ENV), social impact (SOC), economic impact (ECO) and support for the event (SUP). The model demonstrated an adequate fit to the data (χ
2(378) = 1158.76,
p < 0.001; CFI = 0.93; TLI = 0.92; RMSEA = 0.07, 90% CI [0.06, 0.07]; SRMR = 0.05). The standardised factor loadings were all statistically significant and above the recommended thresholds, ranging from 0.66 to 0.92, indicating an adequate relationship between the observed items and their respective latent constructs. Furthermore, the composite reliability (CR) values exceeded the 0.7 threshold for all factors (0.90 to 0.93), and the average variance extracted (AVE) exceeded the reference value of 0.50 in all cases (0.59 to 0.75), which supports the convergent validity of the measurement model.
Discriminant validity was assessed using the heterotrait–monotrait (HTMT) criterion (
Table 5). All the values obtained were below the threshold of 0.85, confirming that the five constructs analysed are empirically distinguishable from one another.
In addition, the factor invariance of the model across genders was assessed using a multi-group confirmatory factor analysis. The model showed an acceptable fit at the configural level (CFI = 0.92; RMSEA = 0.07), and the progressive imposition of restrictions on factor loadings and intercepts did not result in a significant deterioration in the model’s fit at the metric (CFI = 0.92; RMSEA = 0.07) and scalar (CFI = 0.92; RMSEA = 0.07) levels. In accordance with the criteria proposed by [
59], the changes observed in the fit indices were minimal and the indicators remained within the thresholds established in the methodological process, which supports the configural, metric and scalar invariance of the model across genders. Therefore, the results of the confirmatory factor analysis provide an affirmative answer to RQ3, supporting the existence of a multidimensional structure of perceived impact and a differentiated construct of support. As an additional robustness analysis, it was found that this structure is equivalent between men and women.
4.4. Answer to Research Question 4
To answer RQ4, a segmentation analysis was carried out based on the five previously validated dimensions of perception (tourism and sport impact, environmental impact, social impact, economic impact and support), which had been standardised using Z-scores (
Table 6).
Hierarchical cluster analysis using Ward’s method was used to examine the clustering structure of the cases. The agglomeration history showed a marked increase in the agglomeration coefficient when moving from three to two clusters (from 1649.75 to 2675.00), which was substantially higher than that observed in the previous merger from four to three clusters (from 1403.22 to 1649.75). This pattern provided empirical support for selecting a three-cluster solution.
Based on this solution, a non-hierarchical k-means cluster analysis was applied, which converged after 12 iterations. The resulting solution comprised Cluster 1 (n = 171; 31.9%), Cluster 2 (n = 319; 59.5%) and Cluster 3 (n = 46; 8.6%). Descriptive analysis of variance showed that the five dimensions used for segmentation exhibited statistically significant differences between the clusters (p < 0.001 in all cases), providing further evidence of differentiation between the identified profiles. Finally, the socio-demographic profiling of the clusters using contingency tables found no statistically significant associations between cluster membership and socio-demographic variables, such as age (χ2(6) = 8.80, p = 0.185; V = 0.09), educational attainment (χ2(2) = 3.45, p = 0.178; V = 0.08), occupation (χ2(6) = 9.02, p = 0.172; V = 0.09) or income level (χ2(8) = 5.41, p = 0.713; V = 0.07). However, a statistically significant association was found with gender (χ2(2) = 12.41, p = 0.002), with a small effect size (Cramer’s V = 0.15).
Cluster 1, termed “Supporters”, accounts for 31.90% of the sample (n = 171) and is characterised by scores slightly below the mean on the dimensions of perceived inclusive-sporting, social and economic impact, alongside values close to the mean for environmental impact in terms of standardised scores. Perceptions of social (SOC = 4.73) and economic (ECO = 4.34) impact are above the midpoint of the scale. However, this group shows a moderately high willingness to support the event (SUP = 5.57). From a sociodemographic perspective, the cluster tends to comprise a higher proportion of men (36.73% of the total number of men) than women (28.39% of the total).
Cluster 2, labelled “Enthusiasts”, constitutes the largest segment of the sample (59.52%, n = 319) and has the highest scores relative to the centroids across all dimensions of perceived impact, as well as in support for the event, reflecting a clearly positive assessment of the race. It shares with the “Supporters” a low perception of negative environmental impacts, slightly below the midpoint of the scale. This cluster comprises a higher proportion of women (65.48% of the total) compared with the other segments.
Cluster 3, termed “Detractors”, represents 8.58% of the sample (n = 46) and is characterised by scores clearly below the mean across all dimensions of impact perception and support, indicating a predominantly negative assessment of the event. This segment constitutes the smallest group in the sample and has a lower relative proportion of women.
Overall, the results addressing RQ4 provide evidence of differentiated segments in terms of perceived impact and support, as well as a gender bias in membership of these segments, evidenced by the over-representation of women in the “Enthusiasts” cluster and of men in the “Supporters” cluster. This pattern is consistent with a structure of impact perception aligned with the Triple Bottom Line and with the tenets of Social Exchange Theory and Social Representations Theory.